Not correct. A commercial pilot certificate lets you fly for non-scheduled operations - think charter flights, on-demand sightseeing flights, powerline and pipeline inspections. "Uber for the sky" would be exactly this - non-scheduled on-demand operations ("Part 135"). An Airline Transport Certificate is required for scheduled operations - which includes not just Delta, American etc but also FedEx and UPS who fly under Part 121.
It's not based on whether it's passengers or cargo.
I absolutely hate this answer. I graduated from a school (Michigan) that had both EE, CE, and CS. That was 2004 (later went to Carnegie Mellon to do a Ph.D. cause it seemed like the "next thing"). Today I work for a large bay area software company and recruit from UM, amongst a number of other universities. In general, having done a few hundred interviews, I find that CE students (or CS students who tried to go the hardware route within the CS program) are a jack of all trades and master of none. They may have a decent understanding of computer architecture but since that was their major design class, they opted not to take neither an O/S nor a compilers class, so they really have no idea of what is running on top of that simplistic CPU they designed. They didn't take a software engineering class (sure, they did the intro programming but they have no idea what singletons or factories are) so their code sucks. They didn't do VLSI so they don't really know how that Verilog they wrote translates into silicon. They've heard of timing constraints but have no idea how to actually place/route so as to alleviate this.
I could hire one of these people and maybe they'd be able to figure things out. Or, I could hire someone who actually chose a specialty and attained something approaching mastery (relative for a student) of some area. Someone whom I'd actually trust to write good code, who understands threading, who understands the performance implications of what they're writing... or I could hire that guy who knows just enough in a lot of areas to be dangerous.
Can you speak to these parts? Maybe. Would I trust you to work on it? Probably not. Want to prove me wrong? Send me a message and a resume and I'll be happy to set up an interview.
The reality of DRM is that, absent having a TPM that enforces some sort of software integrity that reasonably ensures that the player is sending the video to a trusted display (TPM validating OS validating player software validating HDCP connection), you're going to be stuck with some security-by-obscurity closed source components, or "plugins". It's unfortunate but I can't honestly see a way around that without much larger changes (like trusted computing, but in a slightly less evil implementation hopefully). The "better alternative" to native apps then becomes allowing DRM to be done in the browser in the least intrusive manner possible -- that is, use as much of the browser's code as possible and have the plugin footprint be as small as possible. Today Flash and Silverlight are used not just for DRM but for the entire player application, ideally the player application could be mostly in HTML and using the browser's stack as much as possible, calling out to the DRM module only for either decryption or saying "Please composite this decrypted stream into that div".
It's so tempting to just sit in the corner and say "DRM is evil, we don't want to taint the web with it" but unfortunately, as is often the case in the real world, we don't get to make decisions in isolation of their consequences. DRM on the web is already a reality, largely using Flash or Silverlight (see e.g. Hulu, Netflix). However, both of these platforms face problems -- Silverlight in particular seems to have a rather uncertain future, Flash availability on tablets and mobile in general is largely non-existant. The poster asks us to "please use other applications when necessary" - is this really a good answer? That is going to lead to even less interoperability, and I would argue it hurts the web at a time when it's already fighting a serious battle against native apps that generally offer developers better control (of UI, no random GC pauses, actual threading models, etc). It's easy to say "DRM will harm the web", it's a bit harder to foresee what the eventualities of telling people "please go away and use native apps" are.
I expect this is likely not going to be a popular response, but in short please realize that this is not as simple as saying "DRM is bad". Yes, DRM sucks but I'd argue that in the long run, having a hobbled web platform losing out to native apps (see e.g. iOS) is going to suck more.
I respectfully disagree. I think the career experiences you gain after university can take you in many different directions, and with CS especially you have options in many different fields (far more opportunity to move around from CS to something else than from Aero to CS). Straight out of college, the Aero or MechE degree isn't going to get you a job as a software engineer at Google or Facebook. It may get you a job "programming" somewhere obscure, but you're going to be hard pressed to find a top-tier employer hiring a software engineer these days with a person who is a recent graduate and who didn't do CS. (I also believe there's a difference between computer science, software engineering, and programming, but that's probably a long enough rant for a different post.) The computer scientist could easily get a job at Northrop or Boeing working on embedded control systems for aeronautical / space systems. If they build the systems that model these components and work closely enough in the field, perhaps that bleeds over into an actual aero design job, but to be frank I'd say the chance of that happening is about the same as someone with an aero or MechE background landing a job as a software engineer at a top-tier company. However, the CS person has a much wider swath of potential employers and industries than the Aero person.
Yeah, because I always read forums AFTER I buy something and BEFORE I rip open the box. Come on, when was the last time you bought something, and then did research on it? Me, I figure out whether I want something or not (may involve research, may involve impulses), buy it, and then immediately rip it open and install it.
Actually, neither your SSI card nor your birth certificate are used to establish identity when verifying employment eligibility. Sorry to burst your bubble, but if you actually look at form I-9 you will see that both of those are under List C, which establishes employment eligibility, NOT identity. Documents that establish identity are from lists A or B. (For some reason, a voter registration card establishes identity - I never understood this, because in every single jurisdiction I've ever lived a voter registration card is something that looks like it just came out of a color laser printer on card stock with absolutely no security features whatsoever.) Either way, a birth certificate or social security card merely establishes that the person identified by the document from list B is eligible to work in the U.S., they do not establish the identity of the person - that's the function of the List B documents. (While those documents under List A, such as passports, satisfy both identity and employment eligibility requirements.)
Well, I live in California and here after about the first $40k income is taxed at 9.3%. Not to mention the 8.25% sales tax, the high gas taxes, etc. And before that I lived in DC, where the income tax for anything over (roughly) $30k was also 9.3%, and we had a 10% tax on restaurants (trying to sock it to the tourists, also hurt the locals). So while I'm glad for you that IL only charges 3%, that really doesn't make me feel any better or make the GP's point any less valid.
When I first read the title, I thought they meant WETA as in the PBS station in Washington, D.C. that is responsible for a good chunk of the nation's public tv programs. I too wondered why they put "Weta" from "Weta Workshops" in all caps.
The distribution of the sum will be normally distributed by the CLT, but that turns out to be absolutely useless for modeling. That's why people usually use fractals to generate reasonable datasets to do modeling when it comes to disk / network traffic. (i.e. use something like an 80/20 multifractal). The sum might tell you how much bandwidth you can expect to need in a month, but at any given time? Or for any real modeling purposes? no way. Fractals are very reasonable for simulating network traffic, they have similar propreties (i.e. self similarity at different granularity). So, in short, while you are correct that the CLT does apply to the sum, it's pretty useless in reality.
The third "good" participant is usually Carol, yes. However, the "attacker" is typically "Eve", even if there is no Carol or Dave. The grandparent was talking about a third person wanting to interfere with the affairs of Alice and Bob (split them up), hence this third person is an attacker and not a "good person", hence "Eve" is more appropriate than "Carol".
Suprisingly, Wikipedia actually has an article on Alice/Bob/Eve:
Not to be too nit-picky, but usually when talking about encryption, the parties are Alice and Bob (the two legitimate users), and Eve (the person who is either 'evil' or 'eavesdropping'). I don't think I've ever heard 'Cathy' used as one of the parties...
There's a problem with your argument... you say that if one pays now, one can start saving up for next year now and wind up with the same amount of interest gain. The problem is that you assume the two are exclusive.
There's no reason I can't start saving up now *and still* keep the $10k in the bank until 4/17. Assuming you're making some non-zero return off of your money, it hurts you to give it away any sooner than you have to. No matter when you start saving for next years taxes (or whatever other confounding factors you care to throw in), you are going to lose two months' interest on $10k by filing in February rather than April, and although it's not going to make a huge difference, it's rather pointless to throw that money away.
I think (or at least hope) that you misinterpreted the GP. Let's say, for instance, that it was determined that the botnets were being controlled from Africa (hopefully you could narrow this down a bit more, but for the point of illustration, bear with me). In 2002, there were only 6 IXPs in all of Africa - today, I think it is more, but it's not huge. If you bombed these six sites, you would essentially take Africa offline. Now I certainly don't mean to justify such an action, or say that it is appropriate or would be even effective, all I'm saying is that I think the GP meant that by bombing a few sites, you are basically doing what you call "cutting all connections" (aka "bombing a nation's infrastructure").
Actually, I think you are the one who misunderstands.
The program is basically this: If you bought a retail version of Windows Vista Ultimate, you can buy two additional upgrade licenses for $50 each. These upgrade licenses are for Windows Vista Home Premium - i.e. you don't get two more Ultimate licenses, you get 2 home premium upg licenses. Hence the bit about home premium.
If you bought Vista Ultimate, and have a 64-bit CPU, there's an even easier way... install Vista Ultimate 64-bit version.
The 64-bit DVD booted fine and let me do a clean install from the upgrade CD. I typed in my ugprade CD key, hit next, and it prompted me to accept the license. (The 32-bit CD would not let me do a clean install, I did try that out. It said I had to start the process from within windows.)
While it's true that Common Law is the basis of our legal system, most of the applicable laws when it comes to computer crime are much newer than Common Law. I.e. wiretapping (either phones or email) is not something for which common law is cited, but rather telephone acts from the early 1900s. As such, in the area of gathering computer records, I think it's a safe assumption that there may be very substantive differences between the laws of the US and the UK, making it not such a stupid question to ask. Of course, IANAL:-)
Except that this article isn't about sun using Itanium, it's about sun using Intel's Xeon x86 line. Itanium (IA-64) is a totally different (and now mostly defunct) line.
XML is not for humans. "Assess the validity of the document"??? There's no way in hell that I could give an XML document to the boss down the hall and expect any sort of magic understanding due to there being a few tags. Frankly, a printout of SELECT * is likely just as readable for most cases where a human has any hope. (And I'm guessing you didn't mean validity in the sense of 'Is this XML document valid / standards compliant', because humans definitely suck at that.)
XML's primary purpose is "to facilitate the sharing of data across different information systems, particularly systems connected via the Internet". It was designed to also be human readable, but that was not its primary focus, and I think that's what really screwed it over.
"So what if the machines have to read more characters?" Are you mad man? Let's take our favorite example, the stock quote. I want to send a symbol (typically 4 characters, so let's just say an average of 5 bytes assuming it's null-terminated string) and get back its price (a double, typically 8 bytes, although for stocks since you don't want wierd rounding errors it's probably better to represent it as an integral number of cents, so just use an int64 - 8 bytes). That's an average of 13 bytes. Rather than just opening a socket connection and sending 5 bytes, receiving 8, I send some huge SOAP message that is a request, and get a huge SOAP message back. This is on the order of hundreds of bytes... but let's imagine that somehow we get rid of all of the message, and somehow just magically open a socket and send some XML.
What I'm now sending, even in the best case, is: <stock>.DJIA</stock>. That's around 20 characters, instead of four. My response: <stock> <symbol>DJIA</symbol> <value>12463.87</value> </stock>
Heck, let's ignore all the tags, and just look at "12463.87". I now have to parse out the number, I have to call something like atof (which, by the way, is much worse than something like htonl), and that's totally ignoring the bloat in size for larger numbers, or numbers with a large number of digits (be they either large, or with a large number of digits after the decimal place).
It's just bloated, and it sucks up CPU resources and bandwidth. And please don't tell me that either are cheap - because if you really believe that, then you can pay my monthly bill. I'd gladly sign it over to you.
Does it tell you how to get someone else's public key? Does it walk you through how finding out if the key on pgpkeys.mit.edu is really *my* key or someone else's? And please, do tell me how the hell I'm supposed to get the public key for my bank - any of the five I have a relationship with. Try calling up Bank of America and telling them that you want to send them an email about your account and need their PGP key. If you're lucky you might get someone who has a clue after five transfers, who will just tell you that "Sorry, this is not supported." That's if you're lucky. Now try to get me a key for DeutscheBank. Or, if you really want an exercise in futility, try to tell me how to get a key for Bank of .
Encryption is all well and good, and if you look on pgpkeys.mit.edu you will find my key. I drank the kool-aid a long time ago, but I certainly don't consider encrypted email to be a solved problem. Keyservers, as they are today, are basically a hack. There's no guarantee that you have the correct key. Sure, we could start reading fingerprints and hashes to each other over the phone, but that's far from ideal, and still doesn't solve the problem that if Alice doesn't already know Bob, calling who she believes to be Bob is really not doing all that much to verify any sort of real-world identity if she found Bob's phone number online (the same place she found Bob's key).
The fact that there's a "help" topic does not mean it's a solved problem.
Have you ever tried using this for nontrivial examples? I must confess to being quite fed up with the whole thing. Support for anything beyond the basics seems to vary greatly from library to library, platform to platform, language to language. Axis is probably the best choice for Java, but it's rather lacking when it comes to commercial support, which is important for some people. For C/C++ you're more or less screwed. Gsoap works (for the most part), but it produces the most god-awful stubs I've ever seen. The library that comes as part of Visual Studio (for.Net I believe) either doesn't support MIME or DIME attachments, I don't recall which. There just seem to be too many problems for me to actually bother to use it.
In my opinion, at this point it's just a mess, and for anything beyond the complexity of the stock-quote example I look to other technologies. I, for one, shed no tears at the end of this honeymoon.
(And am I the only one that cringes at using SOAP messages (or XML in general) for something that's supposed to be a machine-to-machine interaction? If you're going to write a new standard, why not write something more efficient?
I call BS. In this year's rankings, DC is #19, behind cities like Memphis, Trenton, and Kansas City. DC has improved greatly since the handgun ban was passed.
As for the murder rate - it definitely does not drop immediately once you cross the border. I'll grant you that the Virginia border takes you into reasonable areas for the most part, but if you go across the Maryland border into PG County, don't tell me that you're going somewhere safe. Frankly, much of the blight is being pushed out of DC into VA and MD due simply to the increasing cost of downtown real-estate. Ten years ago I would never have considered living in Southwest, but now the area is undergoing massive investment, and in my time living there I never had any problems.
Actually, it is bad. See my previous replies as to why you would want to allow mixed sets (hello.jp), plus the fact that you don't need to leave the CJK sets to get roman characters, as they're duplicated in multiple sets all over the place in Unicode.
Slashdot Mirror
Not correct. A commercial pilot certificate lets you fly for non-scheduled operations - think charter flights, on-demand sightseeing flights, powerline and pipeline inspections. "Uber for the sky" would be exactly this - non-scheduled on-demand operations ("Part 135"). An Airline Transport Certificate is required for scheduled operations - which includes not just Delta, American etc but also FedEx and UPS who fly under Part 121.
It's not based on whether it's passengers or cargo.
I absolutely hate this answer. I graduated from a school (Michigan) that had both EE, CE, and CS. That was 2004 (later went to Carnegie Mellon to do a Ph.D. cause it seemed like the "next thing"). Today I work for a large bay area software company and recruit from UM, amongst a number of other universities. In general, having done a few hundred interviews, I find that CE students (or CS students who tried to go the hardware route within the CS program) are a jack of all trades and master of none. They may have a decent understanding of computer architecture but since that was their major design class, they opted not to take neither an O/S nor a compilers class, so they really have no idea of what is running on top of that simplistic CPU they designed. They didn't take a software engineering class (sure, they did the intro programming but they have no idea what singletons or factories are) so their code sucks. They didn't do VLSI so they don't really know how that Verilog they wrote translates into silicon. They've heard of timing constraints but have no idea how to actually place/route so as to alleviate this.
I could hire one of these people and maybe they'd be able to figure things out. Or, I could hire someone who actually chose a specialty and attained something approaching mastery (relative for a student) of some area. Someone whom I'd actually trust to write good code, who understands threading, who understands the performance implications of what they're writing... or I could hire that guy who knows just enough in a lot of areas to be dangerous.
Can you speak to these parts? Maybe. Would I trust you to work on it? Probably not. Want to prove me wrong? Send me a message and a resume and I'll be happy to set up an interview.
The reality of DRM is that, absent having a TPM that enforces some sort of software integrity that reasonably ensures that the player is sending the video to a trusted display (TPM validating OS validating player software validating HDCP connection), you're going to be stuck with some security-by-obscurity closed source components, or "plugins". It's unfortunate but I can't honestly see a way around that without much larger changes (like trusted computing, but in a slightly less evil implementation hopefully). The "better alternative" to native apps then becomes allowing DRM to be done in the browser in the least intrusive manner possible -- that is, use as much of the browser's code as possible and have the plugin footprint be as small as possible. Today Flash and Silverlight are used not just for DRM but for the entire player application, ideally the player application could be mostly in HTML and using the browser's stack as much as possible, calling out to the DRM module only for either decryption or saying "Please composite this decrypted stream into that div".
It's so tempting to just sit in the corner and say "DRM is evil, we don't want to taint the web with it" but unfortunately, as is often the case in the real world, we don't get to make decisions in isolation of their consequences. DRM on the web is already a reality, largely using Flash or Silverlight (see e.g. Hulu, Netflix). However, both of these platforms face problems -- Silverlight in particular seems to have a rather uncertain future, Flash availability on tablets and mobile in general is largely non-existant. The poster asks us to "please use other applications when necessary" - is this really a good answer? That is going to lead to even less interoperability, and I would argue it hurts the web at a time when it's already fighting a serious battle against native apps that generally offer developers better control (of UI, no random GC pauses, actual threading models, etc). It's easy to say "DRM will harm the web", it's a bit harder to foresee what the eventualities of telling people "please go away and use native apps" are.
I expect this is likely not going to be a popular response, but in short please realize that this is not as simple as saying "DRM is bad". Yes, DRM sucks but I'd argue that in the long run, having a hobbled web platform losing out to native apps (see e.g. iOS) is going to suck more.
I respectfully disagree. I think the career experiences you gain after university can take you in many different directions, and with CS especially you have options in many different fields (far more opportunity to move around from CS to something else than from Aero to CS). Straight out of college, the Aero or MechE degree isn't going to get you a job as a software engineer at Google or Facebook. It may get you a job "programming" somewhere obscure, but you're going to be hard pressed to find a top-tier employer hiring a software engineer these days with a person who is a recent graduate and who didn't do CS. (I also believe there's a difference between computer science, software engineering, and programming, but that's probably a long enough rant for a different post.) The computer scientist could easily get a job at Northrop or Boeing working on embedded control systems for aeronautical / space systems. If they build the systems that model these components and work closely enough in the field, perhaps that bleeds over into an actual aero design job, but to be frank I'd say the chance of that happening is about the same as someone with an aero or MechE background landing a job as a software engineer at a top-tier company. However, the CS person has a much wider swath of potential employers and industries than the Aero person.
Yeah, because I always read forums AFTER I buy something and BEFORE I rip open the box. Come on, when was the last time you bought something, and then did research on it? Me, I figure out whether I want something or not (may involve research, may involve impulses), buy it, and then immediately rip it open and install it.
Actually, neither your SSI card nor your birth certificate are used to establish identity when verifying employment eligibility. Sorry to burst your bubble, but if you actually look at form I-9 you will see that both of those are under List C, which establishes employment eligibility, NOT identity. Documents that establish identity are from lists A or B. (For some reason, a voter registration card establishes identity - I never understood this, because in every single jurisdiction I've ever lived a voter registration card is something that looks like it just came out of a color laser printer on card stock with absolutely no security features whatsoever.) Either way, a birth certificate or social security card merely establishes that the person identified by the document from list B is eligible to work in the U.S., they do not establish the identity of the person - that's the function of the List B documents. (While those documents under List A, such as passports, satisfy both identity and employment eligibility requirements.)
Well, I live in California and here after about the first $40k income is taxed at 9.3%. Not to mention the 8.25% sales tax, the high gas taxes, etc. And before that I lived in DC, where the income tax for anything over (roughly) $30k was also 9.3%, and we had a 10% tax on restaurants (trying to sock it to the tourists, also hurt the locals). So while I'm glad for you that IL only charges 3%, that really doesn't make me feel any better or make the GP's point any less valid.
Oh how I wish you were right in your estimate. £475,000 is about US$964,867. (Yes, it's now worse than 2:1).
When I first read the title, I thought they meant WETA as in the PBS station in Washington, D.C. that is responsible for a good chunk of the nation's public tv programs. I too wondered why they put "Weta" from "Weta Workshops" in all caps.
The distribution of the sum will be normally distributed by the CLT, but that turns out to be absolutely useless for modeling. That's why people usually use fractals to generate reasonable datasets to do modeling when it comes to disk / network traffic. (i.e. use something like an 80/20 multifractal). The sum might tell you how much bandwidth you can expect to need in a month, but at any given time? Or for any real modeling purposes? no way. Fractals are very reasonable for simulating network traffic, they have similar propreties (i.e. self similarity at different granularity). So, in short, while you are correct that the CLT does apply to the sum, it's pretty useless in reality.
How much can it lift? Summary says 135kb (500lbs), but 135kg would be roughly 300lbs... I'm confused.
The third "good" participant is usually Carol, yes. However, the "attacker" is typically "Eve", even if there is no Carol or Dave. The grandparent was talking about a third person wanting to interfere with the affairs of Alice and Bob (split them up), hence this third person is an attacker and not a "good person", hence "Eve" is more appropriate than "Carol".
Suprisingly, Wikipedia actually has an article on Alice/Bob/Eve:
http://en.wikipedia.org/wiki/Alice_and_Bob
Not to be too nit-picky, but usually when talking about encryption, the parties are Alice and Bob (the two legitimate users), and Eve (the person who is either 'evil' or 'eavesdropping'). I don't think I've ever heard 'Cathy' used as one of the parties...
There's a problem with your argument... you say that if one pays now, one can start saving up for next year now and wind up with the same amount of interest gain. The problem is that you assume the two are exclusive.
There's no reason I can't start saving up now *and still* keep the $10k in the bank until 4/17. Assuming you're making some non-zero return off of your money, it hurts you to give it away any sooner than you have to. No matter when you start saving for next years taxes (or whatever other confounding factors you care to throw in), you are going to lose two months' interest on $10k by filing in February rather than April, and although it's not going to make a huge difference, it's rather pointless to throw that money away.
I think (or at least hope) that you misinterpreted the GP. Let's say, for instance, that it was determined that the botnets were being controlled from Africa (hopefully you could narrow this down a bit more, but for the point of illustration, bear with me). In 2002, there were only 6 IXPs in all of Africa - today, I think it is more, but it's not huge. If you bombed these six sites, you would essentially take Africa offline. Now I certainly don't mean to justify such an action, or say that it is appropriate or would be even effective, all I'm saying is that I think the GP meant that by bombing a few sites, you are basically doing what you call "cutting all connections" (aka "bombing a nation's infrastructure").
Actually, I think you are the one who misunderstands.
The program is basically this: If you bought a retail version of Windows Vista Ultimate, you can buy two additional upgrade licenses for $50 each. These upgrade licenses are for Windows Vista Home Premium - i.e. you don't get two more Ultimate licenses, you get 2 home premium upg licenses. Hence the bit about home premium.
If you bought Vista Ultimate, and have a 64-bit CPU, there's an even easier way... install Vista Ultimate 64-bit version.
The 64-bit DVD booted fine and let me do a clean install from the upgrade CD. I typed in my ugprade CD key, hit next, and it prompted me to accept the license. (The 32-bit CD would not let me do a clean install, I did try that out. It said I had to start the process from within windows.)
While it's true that Common Law is the basis of our legal system, most of the applicable laws when it comes to computer crime are much newer than Common Law. I.e. wiretapping (either phones or email) is not something for which common law is cited, but rather telephone acts from the early 1900s. As such, in the area of gathering computer records, I think it's a safe assumption that there may be very substantive differences between the laws of the US and the UK, making it not such a stupid question to ask. Of course, IANAL :-)
Except that this article isn't about sun using Itanium, it's about sun using Intel's Xeon x86 line. Itanium (IA-64) is a totally different (and now mostly defunct) line.
XML is not for humans. "Assess the validity of the document"??? There's no way in hell that I could give an XML document to the boss down the hall and expect any sort of magic understanding due to there being a few tags. Frankly, a printout of SELECT * is likely just as readable for most cases where a human has any hope. (And I'm guessing you didn't mean validity in the sense of 'Is this XML document valid / standards compliant', because humans definitely suck at that.)
XML's primary purpose is "to facilitate the sharing of data across different information systems, particularly systems connected via the Internet". It was designed to also be human readable, but that was not its primary focus, and I think that's what really screwed it over.
"So what if the machines have to read more characters?" Are you mad man? Let's take our favorite example, the stock quote. I want to send a symbol (typically 4 characters, so let's just say an average of 5 bytes assuming it's null-terminated string) and get back its price (a double, typically 8 bytes, although for stocks since you don't want wierd rounding errors it's probably better to represent it as an integral number of cents, so just use an int64 - 8 bytes). That's an average of 13 bytes. Rather than just opening a socket connection and sending 5 bytes, receiving 8, I send some huge SOAP message that is a request, and get a huge SOAP message back. This is on the order of hundreds of bytes... but let's imagine that somehow we get rid of all of the message, and somehow just magically open a socket and send some XML.
What I'm now sending, even in the best case, is:
<stock>.DJIA</stock>. That's around 20 characters, instead of four.
My response:
<stock>
<symbol>DJIA</symbol>
<value>12463.87</value>
</stock>
Heck, let's ignore all the tags, and just look at "12463.87". I now have to parse out the number, I have to call something like atof (which, by the way, is much worse than something like htonl), and that's totally ignoring the bloat in size for larger numbers, or numbers with a large number of digits (be they either large, or with a large number of digits after the decimal place).
It's just bloated, and it sucks up CPU resources and bandwidth. And please don't tell me that either are cheap - because if you really believe that, then you can pay my monthly bill. I'd gladly sign it over to you.
Does it tell you how to get someone else's public key? Does it walk you through how finding out if the key on pgpkeys.mit.edu is really *my* key or someone else's? And please, do tell me how the hell I'm supposed to get the public key for my bank - any of the five I have a relationship with. Try calling up Bank of America and telling them that you want to send them an email about your account and need their PGP key. If you're lucky you might get someone who has a clue after five transfers, who will just tell you that "Sorry, this is not supported." That's if you're lucky. Now try to get me a key for DeutscheBank. Or, if you really want an exercise in futility, try to tell me how to get a key for Bank of .
Encryption is all well and good, and if you look on pgpkeys.mit.edu you will find my key. I drank the kool-aid a long time ago, but I certainly don't consider encrypted email to be a solved problem. Keyservers, as they are today, are basically a hack. There's no guarantee that you have the correct key. Sure, we could start reading fingerprints and hashes to each other over the phone, but that's far from ideal, and still doesn't solve the problem that if Alice doesn't already know Bob, calling who she believes to be Bob is really not doing all that much to verify any sort of real-world identity if she found Bob's phone number online (the same place she found Bob's key).
The fact that there's a "help" topic does not mean it's a solved problem.
Have you ever tried using this for nontrivial examples? I must confess to being quite fed up with the whole thing. Support for anything beyond the basics seems to vary greatly from library to library, platform to platform, language to language. Axis is probably the best choice for Java, but it's rather lacking when it comes to commercial support, which is important for some people. For C/C++ you're more or less screwed. Gsoap works (for the most part), but it produces the most god-awful stubs I've ever seen. The library that comes as part of Visual Studio (for .Net I believe) either doesn't support MIME or DIME attachments, I don't recall which. There just seem to be too many problems for me to actually bother to use it.
In my opinion, at this point it's just a mess, and for anything beyond the complexity of the stock-quote example I look to other technologies. I, for one, shed no tears at the end of this honeymoon.
(And am I the only one that cringes at using SOAP messages (or XML in general) for something that's supposed to be a machine-to-machine interaction? If you're going to write a new standard, why not write something more efficient?
I call BS. In this year's rankings, DC is #19, behind cities like Memphis, Trenton, and Kansas City. DC has improved greatly since the handgun ban was passed.
As for the murder rate - it definitely does not drop immediately once you cross the border. I'll grant you that the Virginia border takes you into reasonable areas for the most part, but if you go across the Maryland border into PG County, don't tell me that you're going somewhere safe. Frankly, much of the blight is being pushed out of DC into VA and MD due simply to the increasing cost of downtown real-estate. Ten years ago I would never have considered living in Southwest, but now the area is undergoing massive investment, and in my time living there I never had any problems.
Actually, it is bad. See my previous replies as to why you would want to allow mixed sets (hello.jp), plus the fact that you don't need to leave the CJK sets to get roman characters, as they're duplicated in multiple sets all over the place in Unicode.